Back
Technology

Google Launches AI-Powered 'Auto Browse' Feature for Chrome in US Preview

Generated on: Last updated: 6 sources
View source

Google Launches "Auto Browse" AI Feature for Chrome

Google has introduced a new "Auto Browse" feature for its Chrome desktop browser, powered by its Gemini 3 generative AI model. The tool is designed to automate multi-step online tasks such as shopping, form-filling, and appointment scheduling. Currently, the feature is available in a preview state exclusively to paid subscribers in the United States.

Feature Overview and Functionality

Auto Browse integrates Google's Gemini AI assistant directly into the Chrome browser via a right-hand sidebar. Users can initiate tasks by providing instructions to Gemini, which then operates independently within a browser tab to complete the requested actions.

According to Google, internal testers and demonstrations have shown the feature capable of:

  • Shopping online, including searching for items, adding them to a cart, and applying discount codes.
  • Filling out online forms using data from provided documents.
  • Scheduling appointments and gathering documents.
  • Managing bills, subscriptions, and expense reports.
  • Expediting processes like driver's license renewals.

During a task, Gemini indicates "Task started" and interacts with web pages by clicking, typing, and scrolling.

A "Take over task" button is present, allowing users to interrupt and manually control the browser at any point.

Availability and Access

The Auto Browse feature is currently in a preview phase with limited availability:

  • Region: United States only.
  • Access: Available to subscribers of Google’s paid AI plans: AI Pro ($20/month) and AI Ultra ($250/month).
  • Usage Caps: Subscribers have daily request limits: 20 tasks for AI Pro and 200 tasks for AI Ultra.
  • Launch: Users can activate the feature through an opt-in pop-up or by enabling the 'Let Chrome browse for you' toggle in Chrome settings after accessing the Gemini sidebar.

Google has not specified a timeline for a wider release to free users or other countries.

User Control and Security Considerations

Google includes disclaimers noting that users are responsible for Gemini's actions during tasks and should "take control if needed."

The feature is designed with specific safeguards:

  • User Confirmation: For sensitive actions deemed to carry higher risk—such as completing a purchase, posting on social media, or processing credit card transactions—Auto Browse is designed to require explicit user confirmation for the final step.
  • Reported Performance: Initial attempts by some users to access the feature reportedly resulted in Gemini stating it had "taken over the navigation" without performing actual browser clicks, requiring a browser refresh.

Google and industry observers note potential security concerns with automated AI browsing agents, including:

  • Prompt Injection Attacks: The risk that visiting a malicious website could trick the AI into performing unintended actions.
  • Incorrect Information: The potential for AI to generate or act upon inaccurate data.

Google states it is developing "layered defenses" to address these threats.

Industry Context and Development

The release aligns with a broader industry trend of integrating "agentic" AI functionalities into web browsers, with competitors like Microsoft, OpenAI, and Perplexity offering similar capabilities.

Google is developing infrastructure for Gemini and Auto Browse to integrate with third-party platforms. The company notes that Chrome will support the Universal Commerce Protocol (UCP), an open standard co-developed with companies including Etsy, Shopify, Target, and Wayfair. UCP is designed to facilitate standardized interaction between AI agents and diverse e-commerce websites.